The Evidence Base

Informing Policy in Health, Economics & Well-Being
A collaboration with
USC Dornsife Center for economic and social research

Didn’t you Hear? The Understanding America Study is Here.

With Election Day 2015 just past, it’s less than year to go before America picks a new president – and that means polls, polls and more polls. Pre-Iowa caucuses. Post-Iowa/pre-New Hampshire. Super Tuesday. California’s primary on June 7. Conventions. Debates. Right up until Election Day 2016, on Nov. 8.

USC CESR has developed a strong yet sprightly polling tool – the Understanding America Study (UAS) – featuring a nationwide sample of citizens notable for its breadth and depth, necessary for precise results. Adding upon those strengths, IT experts actively are working with the poll’s guts to coax even more knowledge from the panel of 2,250 households, whose opinions are recorded via online devices once or twice a month.

To date, the UAS hasn’t been used to poll the presidential race, although CESR Director Arie Kapteyn said it’s under consideration. Instead, the topics probed concern matters such as health care and financial literacy.

But the UAS has promise to deliver a crystal-clear snapshot of the race for the White House, if polling does start.

For one, said those who work with the UAS, every time the poll is conducted the respondents are the same people – which means changes in results have a higher confidence of being real changes in these individuals’ minds. “Because you’re asking the same people all the time, you all of a sudden don’t have the effects of asking a new group,” said Bart Orriens, Managing IT Director of CESR.

Polling is difficult work, with its accuracy like a golf ball in shifting winds, dependent upon quirks in how respondents are chosen and give their opinions. In person, phones and online, or online only? If phones, landlines and cell? In which order are the questions asked and the choices presented?

On that last question, the UAS accounts for those hidden biases by having the software – which goes by the name NUBIS – switch up the order of queries and answers. Explained Jillian Wallis, Research Data Administrator with the Schaeffer Center: “People can be affected quite a bit by the order of answer categories – It’s amazing to see the effects (of randomization.)”. Other survey features that can be customized include explainer videos and slider questions (“On a scale of 1 to 100…”).

At the end, CESR researchers are thrilled by what they’re receiving.

“The ability to launch a survey and get the results back within a couple of weeks is amazing,” said Joanne Yoong, Senior Economist at CESR East. “And it’s way better than any other online survey we know of because the sampling is very particular.”

The UAS isn’t just picky; it’s also growing. Kapteyn said the goal, by next summer, is to have 6,000 households participating.

“And they’re really making an effort to recruit minority populations,” Yoong said. “Native American populations, low-income populations – these are populations that typically get under-represented in surveys.”

Yoong and colleagues recently used the UAS while determining just how knowledgeable and prepared Americans are when it comes to their retirement planning. About 45 questions produced the paper What do people know about Social Security? (Spoiler: Alarmingly little.)